The proposed study will evaluate the long-term effectiveness of the Life Skills Training (LST) program for preventing onset of cigarette smoking in a cohort of urban Hispanic students, and will obtain basic knowledge about the family determinants of smoking behavior in this population. In a randomized-blocks design, 30 schools with 50% or more Hispanic students will be randomly assigned to either experimental or control groups. In the experimental group, schools will receive the complete LST program for smoking prevention, which emphasizes social and personal competence skills; in the control group, schools will receive only existing educational activities. Students will be followed over a 4-year period to determine the long-term effectiveness of the program. Within the experimental group, half the schools will be junior high schools, and the LST intervention will begin in 7th grade; half the schools will be intermediate schools, and within this group intervention will begin in the 6th grade and 7th grade for half the schools, and will begin only in 7th grade for the other half of the schools. The primary evaluation question will be addressed by using school means (n = 15 per group) as the unit of analysis. The secondary evaluation question will compare effect of intervention in junior high school with effect of intervention in intermediate school using individual subjects as the unit of analysis (n = 1100 per group). The primary dependent measure wil be enhanced self-reports of cigarette smoking using the "bogus pipeline" technique, supplemented by subsamples of saliva thiocyanate (SCN). The impact of the prevention program on mediating cognitive, attitudinal, and psycholsocial variables will be assessed, and moderator effects of sex, ethnicity, and family characteristics will be analyzed. Additionally, telephone interviews with families of Hispanic students in experimental and control groups (total n = 2,000 families) will be conducted to obtain data on smoking and drinking behavior anong Hispanic families and the relation of these behaviors and other family characteristics (e.g., health attitudes and family stress) to smoking among Hispanic adolescents. Developmental research will be conducted to develop family intervention components that can be added to the school-based prevention program. In conclusion, the project offers the potential of providing an effective approach to smoking prevention with Hispanic adolescents that would reduce cancer risk in this population.